Professional Documents
Culture Documents
Electronic System For Data Acquisition and
Electronic System For Data Acquisition and
Brasília, DF
2018
João Victor Avancini Guimarães
Brasília, DF
2018
João Victor Avancini Guimarães
Electronic System for Data Acquisition and Control of a Automotive Brake
Test Bench/ João Victor Avancini Guimarães. – Brasília, DF, 2018-
114 p. : il. (some colors.) ; 30 cm.
CDU 02:141:005.6
João Victor Avancini Guimarães
Brasília, DF
2018
Acknowledgements
First of all I thank my parents Rita and Carlos and my siblings Pedro, Ana and
Frederico for all their efforts, dedication and support over the years.
To my supervisor Prof. Dr. Evandro Leonardo Silva Teixeira for his patience,
support and teachings given throughout my stay at the University of Brasilia.
To my friends Mairon, Wener and Joel for the comradeship along our stay at
University of Brasília.
To my work colleges during my internship at Autotrac Comércio e Telecomunica-
ções S/A for sharing their experience and knowledge with me.
To the lecturers Julia Peterle, Aline Demuner, Graciela Ramos, Carmen Santos,
Gerardo Pizo, Cristiano Miosso, Gilmar Beserra, Richard Pearl, Josh Reynolds, Eleanor
Baldwin, Steve Hegarty, Wellington Amaral, Sebastién Rondineau and other teachers that
I had the privilege of learning from over the years.
“Science, my lad, is made up of mistakes, but they are mistakes which it is useful to
make, because they lead little by little to the truth.”
(Jules Verne, A Journey to the Center of the Earth)
Abstract
With the evolution of the automotive industry, vehicles come out of the factory each time
with more power and faster top speeds. Along with this enhancement in performance the
concern in effectiviness of brake systems should also increase.
There are already consolidated standards rules for brake system testing, this project is
focused with respect to SAE J2522 regulation that addresses on brakes tests on passenger
vehicles.
This paper aims to study the major factors regarding brake tests and electronic instrumen-
tation in order to develop an electronic instrument system for automation of a testbench
used to perform brake tests. The major output of the project is to design an electronic
system able to measure all the relevant physical data and to control the needed actuators.
IC Integrated Circuit.
I2 C Inter-Integrated Circuit.
SNR Signal-to-Noise-Ratio.
HMI Human-Machine-Interface.
List of symbols
cm 10−2 Meters.
ms 10−3 Seconds.
mv 10−3 Volts.
mA 10−3 Ampere.
1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.1 Purpose of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.2 Text Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2 LITERATURE REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.1 Working principles of disk brake systems . . . . . . . . . . . . . . . . 29
2.2 The SAE J2522 regulation . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.1 Monitored Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.2 Brake Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3 Review of basic Electronic Sensors, Devices and Concepts Relevant
to this thesys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3.1 Rotation Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3.2 Temperature Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3.3 Brake Force Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.4 Vibration Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.5 Instrumentation Amplifier . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.3.5.1 Important features to consider in an amplifier . . . . . . . . . . . . . . . . . . 38
2.3.6 System Controlling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3.6.1 Microcontroller basic characteristics . . . . . . . . . . . . . . . . . . . . . . 40
2.3.6.2 Microcontroller Architecture and Most Important Parts . . . . . . . . . . . . . 40
2.3.6.3 Pulse Width Modulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.3.7 Circuit Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.3.7.1 Low Pass Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.3.7.2 Transient Voltage Supression Diodes (TVS) . . . . . . . . . . . . . . . . . . 44
2.3.7.2.1 TVS diodes principle of operation . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.3.7.2.2 Selecting TVS Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3 METHODOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.1 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Problem Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3 Project Conception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.4 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5 FIRMWARE PROJECT . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.1 Microcontroller programming basics . . . . . . . . . . . . . . . . . . . 85
5.2 Code Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6 EXPERIMENTAL TESTS . . . . . . . . . . . . . . . . . . . . . . . . 89
6.1 Hardware Interfaces Test . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.1.1 Rotation Aquisition Channel Test . . . . . . . . . . . . . . . . . . . . . . 89
6.1.1.1 Configuration and Background . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.1.1.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.1.2 Temperature Acquisition Channels Test . . . . . . . . . . . . . . . . . . . 91
6.1.2.1 Configuration and Background . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.1.2.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.1.2.2.1 Channel 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.1.2.2.2 Channel 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.1.2.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
6.1.3 Brake Force Acquisition Channels Test . . . . . . . . . . . . . . . . . . . . 92
6.1.3.1 Configuration and Background . . . . . . . . . . . . . . . . . . . . . . . . . 92
6.1.3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.1.3.2.1 Channel 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.1.3.2.2 Channel 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.1.3.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.1.4 Analog Speed Reference Output Channels Test . . . . . . . . . . . . . . . 94
6.2 Brake Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.2.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.2.2 Test Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.2.3 Full Stop Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.2.4 Full Stop Tests Results and Discussion . . . . . . . . . . . . . . . . . . . . 101
6.2.5 Frequency Inverter Analog Input Test . . . . . . . . . . . . . . . . . . . . 103
6.2.5.1 Analog Output Duty Cycle Tests . . . . . . . . . . . . . . . . . . . . . . . . 103
6.2.5.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.2.6 Vibration Testes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.2.6.1 Configuration and Background . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.2.7 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.2.7.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
25
1 Introduction
With the advancement of technology, cars are leaving the factory each time with
more power and affordable prices. In 1995 the best-selling car in Brazil (HERNANDES,
2017), (Volkswagen Gol Plus 1.0), had 49.8hp of maximum power and brand-new would
accelerate from nought to 100kph in 22.4 seconds (CNW, 2017b). On the other hand, the
sales champion of 2015 ,Volkswagen Gol 1.0, had 76hp of maximum power and could do
the same task in 13.3 seconds (CNW, 2017a), almost half the time from the previous.
Interesting fact is that even though the latter is 20 years younger, both cars have similar
brake systems: disk brakes in the front wheels and drum ones in the back. This enhance-
ment on vehicle accelerations naturally imply on higher top speeds, which should lead to
a bigger concern in brake systems effectiveness.
During the development of any solution or enhancement on any technology or
product, testing is fundamental. In the matter of brake systems, the fact stated on the
previous sentence can also be considered true. On the process of developing more effective
brake systems, it is important to perform brake tests.
Although brake tests are so important (ABENDROTH, 1985), performing this
tests with full scale vehicles are obviously expensive and this somehow makes extensive
testing unfeasible. Also the time required for each test might be a constrain. The automo-
tive industry has being using testbenchs and controlled environments in order to provide
relevant and reliable information about quality and performance of automotive systems
with lower costs and reduced testing time. Small scale tests do not have the purpose of
fully replacing full scale ones, but the savings in costs and time that they could provide can
be used for mass testing, and this can already somehow show their utility and relevance
(GARDINALLI, 2005).
Evaluating the brake efficiency of a vehicle as a whole, involves a lot of factors, a
small scale test plataform will not provide results that could be used to fully address the
quality of a car break system, instead, it is possible to focus the results on the performance
of individual componentes of the system such as pads, disks and calipers (HALDERMAN;
MITCHELL, 2016), and evaluating the performance of this components is a good start
for evaluating the braking capacity of a brake system.
Government and legal authorities have been creating strict regulations for manu-
factures to follow in order to ensure that cars have a higher standard of safety. Braking
tests have been regulated for some time. An international standard for brake testing is
the regulation SAE J2522 (SAE, 2016), which supplies a description of how break tests
should be conducted for systems used in low weight passengers cars. The brake system
26 Chapter 1 Introduction
is a critical part of an automobile, thanks to this system it is possible to use the latter
under safe conditions both in urban and rural areas. There are some ideal requirements
that a brake system should be able to attend (KAWAGUCHI, 2005):
∙ Reduce the speed of a moving vehicle, increasing the deceleration of the same.
Another point of view should also be considered during this analysis, the manu-
facturers point of view. More effective brake systems means more research and probably
more expensive materials, generating more cost to manufacturers and consequently more
cost to customers. Theoretically this would meant that manufacturers need to balance the
trade-off between quality and cost. However, the point in which this trade-off is setted is
determined by government regulations. Moreover if there was no general regulations each
car manufecturer would have a standard that they judge is sufficient. In Brazil, according
to Resolution N.519 (CONTRAN, 2015) from the National Traffic Council (CONTRAN -
Conselho Nacional de Trânsito), the minimal requirements of perfomance of brake systems
from any vehicle with 750kg or less should meet the following regulations from Brazilian
Association of Technical Standards (ABNT - Associação Brasileira de Normas Técnicas):
NBR 10966-1, NBR 10966-2, NBR 10966-3, NBR 10966-4, NBR 10966-5, NBR 10966-
6, NBR 10966-7 and NBR 16068. Most of these regulations are based in the european
regulation ECE-13/05.
Considering the importance of regulatory standards, the need for brake tests be-
comes even more evident as it is mandatory to ensure that brake-systems will attend
to regulations requirements. Only with extensive testing it is possible to ensure that a
particular system will attend to all standards regarding it’s category of operation.
Making all these considerations, a Break-Testbench can be a useful device for the
automotive industry and research. Considering that this asset would be able to simu-
late similar environmental and operational conditions that a brake system is submitted,
this testbench could allow manufactures and suppliers to avoid expenses in field tests as
they would be able to test different parts of the system in an assisted and controlled
evironments.
the information from the internationl regulation SAE J2522 (SAE, 2016). The system
should be able to perform brake tests by controlling the mechanical parts while acquiring
relevant data with the aid of sensors and transducers.
2 Literature Review
As Figure 1 shows, some of the main components of a typical disk brake system
and their respective functions are (CARPARTS, 2018):
∙ Caliper: Fits over the rotor like a clamp, houses the pads and pistons. This part
does not tend to wear out quickly, usually, damages on this part are caused by using
outworn pads or disks (GOODYEAR, 2018).
∙ Pads: There are two brake pads on each caliper, this is the part that is effectively in
contact with the disk. Pads are used in order to prevent disk wear and caliper wear,
30 Chapter 2 Literature Review
this parts are made with specific materials that provide better perfomance, hence
more friction with less dissipated heat. Pads tend to wear out quickly in a vehicle
and should be checked and replaced quite often.
∙ Piston: Most of brake discs are activated through a hydraulic mechanism, when the
break pedal is pressed the brake fluid goes to the piston and the piston squeezes the
to brake pads against the disk.
∙ Rotor: The disk rotor is the main component of the system, it is directly connected to
the wheel’s shaft. The rotor wears with time, if break pads are replaced accordingly
the rotor tends to last more than the pads. Usually a vehicle will undergo more than
one brake pads replacement before having its disks replaced.
∙ Temperature of the brake pads: It is important during the entire test to be fully
aware of the temperature of the brake system, firstly by the safety factor (there is
a maximum operating temperature for the system), also by the wear of the system
2.2. The SAE J2522 regulation 31
that is tied to the temperature therein. Last but not least is the fact that knowing
the temperature it is possible to carry out tests based on it, being possible to make
tests in known temperature ranges.
∙ Braking force: The pressure to be measured is the pressure that the brake caliper is
applying to the rotor over time, knowing the magnitude of this force means having
control over how much the temperature increases according to braking and how
much the rotor speed decreases as a function of braking.
The system is accelerated to a top speed (Upper Speed Limit), than the brake
actuates with a known force (Brake Force) until the system reaches a certain speed
(Lower Speed Limit), this process is repeated for certain number of times (Snubs)
32 Chapter 2 Literature Review
and after each test the system must rest (Cooldown) until the brakes are cooled enough
regarding a defined cooldown time.
This sensor is widely used in the automotive industry to determine the angular
position of the vehicle’s crankshaft and its rotation speed. There are several types of CKP
sensors, the most common are the variable reluctance type because they have low cost
and good accuracy (SCHROEDER, 2002).
As the gear in question rotates each tooth of the gear aligns with the variable
reluctance sensor, a magnetic flux in the sensor coil changes as the air gap between the
sensor and the gear changes. This change in the magnetic field generates induces a voltage
at the sensor output. This type of sensors have an analog voltage output where amplitude
and frequency vary proportionally to the speed of rotation of a gear, as Figure 4. With
this type of sensor it is possible to extract data of linear velocity, angular velocity and
2.3. Review of basic Electronic Sensors, Devices and Concepts Relevant to this thesys 33
angular position. However, only the angular velocity data (frequency) is important to be
measured for brake tests specified by the regulation (SAE, 2003).
ductor based ICs. It is stated by (NEWTON, 2016) that brake pads from production
passenger cars operate at the temperature range from 100∘ C to 650∘ C. Among the most
common type of temperature transducers, thermocouples operate across the widest tem-
perature range, from -200 ∘ C to 1750 ∘ C (AMETHERM, 2018). Moreover, regulation SAE
J2522 (SAE, 2003) states that temperature measurements for performing specified brake
tests should be done with a Thermocouple. Taking the sum of all this facts, thermocouples
seems to be the reasonable choice for this kind of application.
A thermocouple is a device that has this junction of two different metals and has
a known voltage generated output proportional to the heat tranfer on the junction. In
theory any combination of two different metals could be used, but there are standardized
combinations that produces more stable and predictable voltage outputs (POLLOCK,
1991). This relation between heat transfer and voltage is known not to be linear as Figure
5 shows (E,J,K,T,R,S and B are designators for normalized thermocouples junction types).
Thermocouple actually have two junctions, a hot junction (the one that is sub-
mitted to heat transfers) and a cold junction (also called reference junction). What the
thermocouple really measures is the difference between the temperature of this two junc-
tions, this means that in a hypothetical situation which the hot junction is submmited
to a 100∘ C and the cold junction is submitted to an environmental temperature of 25∘ C,
after thermal equilibrium is reached the thermocouple voltage will be proportional to
a temperature of 75∘ C. Hence the thermocouple will only produce a r̈ealöutput voltage
when the cold junction is submitted to a 0∘ C (in some calibrations procedures the cold
2.3. Review of basic Electronic Sensors, Devices and Concepts Relevant to this thesys 35
There is a wide range of thermocouples available, table 1 shows the most common
thermocouples and there temperature range.
Table 1 – Thermocouples and their operation ranges
Generally, the length variation in a strain gauge is very small and this makes them
very susceptible to measurement errors. As a result, the use of a Wheatstone bridge is
36 Chapter 2 Literature Review
very common, it is formed by four resistive arms and an excitation voltage applied to the
bridge (WINDOW; HOLISTER et al., 1982).
The voltage output 𝑉𝑂 can be obtained from the Equation 2.1, 𝑉𝑂 the load cell
signal output is a differential pair signal.
𝑅3 𝑅2
𝑉𝑂 = − (2.1)
𝑅3 + 𝑅4 𝑅1 + 𝑅2
According to (DILLON; GRIFFEN; WEIHS, 1989) the main advantages and draw-
backs of Load Cells are:
∙ Calibration: load cells are usually already calibrated by manufacturers and vendors.
∙ Size and Weight: load cells are bigger and heavier than most force transducers, for
applications where size and weight is an important requirement they may not be
suitable.
Taking all this characteristics into concern the load cell is the most suitable trans-
ducer for an automotive test environment.
The instrumentation amplifier has two stages, the first stage consists in amplifying
both inputs of a sensor, with the gain of this amplification stage controlled by 𝑅𝑔𝑎𝑖𝑛 in
Figure 10. The second stage consists in taking the difference of the two input signals. If the
differentiation appears before the amplification, noise may be so huge in the input signals
that some signal information might be lost. In the instrumentation amplifier noise and
signal is amplified on the first stage, considering that the noise is similar in both inputs
the differential stage will take out the noise and only output the difference between both
inputs. One advantage of this amplifier is that it has high input impedances, that means
it will not drain significant current from the signal, i.e., it will not interfere with the
measure (THOMSEN et al., 2003). Another advantage is that the gain of this amplifier
can be adjusted with just one resistor (METTINGVANRIJN; PEPER; GRIMBERGEN,
1994).
The gain of the instrumentation amplifier is given by the following 2.3.5 (COUNTS;
KITCHEN, 2006).
(︃ )︃
2·𝑅
𝑉 𝑜𝑢𝑡 = (𝑉2 − 𝑉1 ) · 1 + (2.2)
𝑅𝑔𝑎𝑖𝑛
Something interesting to notice is that if we take out the 𝑅𝑔𝑎𝑖𝑛 (open load), the gain
of the amplifier is equal to one. Besides this advantageous behavior of the instrumentation
amplifier, it is quite hard to make it work with discrete components assembly because
of components imprecision. Hence, it is more practical to work with ICs that ensures
symmetry of between those components. As an example, there is the Texas Instruments
INA118 (INSTRUMENTS, 2000b), which is one of many encapsulated solutions for the
instrumentation amplifier, the schematic of this component is shown in Figure 11.
There are some important factors when choosing an OPAMP in a certain applica-
tion, according to (INC.(NORWOOD, 2011), some of the important features to know in
2.3. Review of basic Electronic Sensors, Devices and Concepts Relevant to this thesys 39
an amplifier are:
∙ Operating Supply Voltage Range: The supply voltage range that can be applied to an
amplifier for which it operates within specifications. Many applications implement
op amp circuits with balanced dual supplies, while other applications for energy
conservation or other reasons, use single-supply.
∙ Supply Current: The current required from the supply voltage to operate the am-
plifier with no load.
At the design of this block of the solution an important thing should be considered,
data acquisition. According to (NATIONAL INSTRUMENTS, 2018) DAQ (Data Acqui-
sition) can be defined as the process of measuring an electric or physical phenomenon,
such as voltage, current, temperature, pressure or sound.
As mentioned in Section 1.1: "The system should be able to perform brake tests by
controlling the mechanical parts while acquiring relevant data with the aid of sensors and
transducers". There is a type of device that is able to perform data acquisition and system
control, the microcontroller.
In order to understand how a typical MCU parts are, Figure 12 show the block
diagram of the ATmega32U4, a MCU manufactured by Microchip.
The most important blocks of this device are:
∙ Timing and Control: This circuit block represents the part of MCU responsible for
maintaing the clock signals and handle time events.
2.3. Review of basic Electronic Sensors, Devices and Concepts Relevant to this thesys 41
∙ Program Flash: This is a non-volatile memory used to store the code, MCUs can
execute code directly from the flash memory.
∙ Port Drivers: This are the device I/Os (inputs/outputs), they can be digital or
analog ports (execpt for outputs which can only be digital).
∙ ADC (Analogic to Digital Converter): MCUs are by nature digital devices, when
acquiring an analog signal, they first make a conversion to a digital value and then
copy this data to a register.
∙ USB 2.0 : Although not being a common feature, some MCUs include an USB (Uni-
versal Serial Bus) port. When interfacing with other USB devices, already having
this feature inside the MCU saves space and costs when integrating the MCU in a
circuit board.
42 Chapter 2 Literature Review
∙ High Speed Timer/PWM : As mentioned in Item 2.3.6.2, some MCUs do not have
analog outputs, the high speed is used to generate a special digital output signal
which will be explained on Section 2.3.6.3.
Duty cycle can be defined mathematically by the Equation 2.3 (JAMES, 2001),
where 𝐷% is the duty cycle in percentage, PW is the pulse width (pulse active time) and
T is the wave period.
𝑃𝑊
𝐷(%) = (2.3)
𝑇
One useful way of using a PWM duty cycle variation is by enconding an analog vol-
tage level proportional into it’s percentage (HOLMES; LIPO, 2003), meaning that 100%
duty cycle would represent maximum voltage amplitude and 0% the minimum voltage.
This means it is possible to extract an analog voltage level by taking a PWM average
level (ALTER, 2008), this is a practical way for designing digital to analog converters.
systems must be protected from possible damage that may be caused by unintentional/ac-
cidental high voltage inputs, surges, etc (MATHIVANAN, 2007).
According to (LITTLELFUSE, 2015b), Voltage Transients are defined as short
duration surges of electrical energy and are the result of the sudden release of energy
previously stored or induced by other means. There are many things that can cause
voltage transients (commonly refered just by transients) and those can be divided in two
groups:
In order to enhance energy efficiency, devices are now operating at lower voltages
(SOUZA et al., 2017). With the miniaturization of electronic components, those have
become even more sensitive to electrical stress (LITTLELFUSE, 2015b).
In automotive environment, many of the supporting electrical components of the
vehicle can generate transients, specially from inductive load switching. When the induc-
tive load is switched off, the collapsing magnetic field is converted into electrical energy
that turns into a transient (LITTLELFUSE, 2015b).
Considering that a signal has a frequency that is lower than the frequency of the
transient it is vulnerable to, using a low pass filter to provide suitable attenuation for that
transient and still letting the signal frequency on the passaband, it is possible to protect
that part of the circuit against this particular transient (STANDLER, 1988).
Transient Voltage Supression Diodes or TVS Diodes are the most popular choice
for protection components in circuit due to their fast response, low clamping voltage and
longevity. Under normal operation TVS diodes are high-impedance devices, interacting
as an open circuit to the protected component, during a transient event the TVS diode
junction provides a low-impedance path for the transient current (RENESAS, 2016a).
There are both uni-directional and bi-directional, the first has an operation curve
similar as the one from a Zener diode, during positive transients the device limits the input
voltage to it’s clamping voltage and during negative transients the spike is clamped to
the diode drop. On the other hand, bi-directional TVS diodes are always reverse biased
during both negative and positive transients. Figures 15 and 16 show respectively the
action of an uni-directional TVS and a bi-directional TVS during transient events.
The Voltage X Current curve of both uni-directional and bi-directional TVS diodes
can be seen repectively on Figures 17 and 18.
2.3. Review of basic Electronic Sensors, Devices and Concepts Relevant to this thesys 45
∙ 𝑉𝐶 (Campling Voltage): the voltage limit that the TVS will allow on the point of
intended protection. This voltage should be slightly lower than the point of intended
protection absolute maximum voltage rating.
∙ V𝑊 𝑀 (Rated Standoff Voltage): this is the maximum voltage in which the TVS still
works as a high-impedance device. Therefore, the first step when selecting a TVS
device is to know the peak voltage at the point of intended protection during normal
operation and then select a TVS device with an appropriate 𝑉𝑊 𝑀 .
∙ P𝑃 𝑃 (Peak Pulse Power): in order to ensure the durability of the TVS device, the
maximum power it needs to dissipate should be known. This P𝑃 𝑃 is calculated
by multipling the TVS device 𝑉𝐶 (Clamping Voltage) and the 𝐼𝑃 𝑃 (Peak Impulse
Current, peak current for the transient event).
47
3 Methodology
In other to achieve the goals of this project the endeavours were divided into three
∙ Temperature of brake pads: During all test it is mandatory to have full knowledge
of the temperature of the break pads, firstly because of security reasons (there is
upper limit for temperature in any system) and also because of the wear of parts
that is related to temperature.
∙ Pressure applied on the disks: Knowing the magnitude of this force means being
able to relate the pressure applied and the deceleration, knowing how the pressure
applied increases the temperature of the pads and evaluate how this promotes wear
of the parts.
∙ Rotation speed: Without knowing how the speed of the rotor varies over time it
would be impossible to determine the acceleration and deceleration rates among
many other issues.
∙ Vibration: As mentioned before this is not mandatory but rather interesting, mea-
suring vibration makes it possible to determine how the extensive use can wear out
the parts and reduce stiffness among other properties. Also it is natural that the
system will vibrate during braking, minimal vibration or too much vibration can
indicate a fault that on the future could damage the system.
Also some other optional requirements were created for making the system more
versatile for future tests, thoose being:
∙ Having a built-in USB to UART transceiver to make the system more practical to
use.
be useless because no possible technical analysis could be done afterwards. The Trans-
ducers/Sensors block was placed within the Brake System Simulator block because the
sensors and transducers will be placed arround the components of this major block.
It is possible to see that each type of transducer will have a specific circuit in order
to perform signal conditioning as better as possible according to the type of transducer
parameters. Also it is possible to see that there will be special circuits to interface the
MCU with digital outputs and inputs alongside with analog outputs. Regarding the analog
outputs, it is also interesting to have a feedback path to the MCU so adjustments can
be made in order to guarantee that the analog outputs voltages are correct according to
the desired quantities. Moreover, a power supply block should be designed to generate a
stabilized voltage to power up all the circuit modules. Last but not least, a communication
port (USB) was placed, through this port the communication with the upper layer of
hardware will be done.
Based on the previous considerations and considering possible future usages for
the designed product, this are the following hardware interfaces.
∙ Sensor Inputs:
– Speed Sensor.
– Temperature Sensor 1.
4.4. Rotation Acquisition Channel 53
– Temperature Sensor 2.
– Brake Force Sensor 1.
– Brake Force Sensor 2.
– Vibration Sensor 1.
∙ Digital Interfaces:
– Output 1.
– Output 2.
– Output 3.
– Input 1.
– Input 2.
– Input 3.
∙ Analog Outputs:
– Output 1.
– Output 2.
∙ Communication Ports:
– USB Port 1.
The conditioning circuit for this project was based on the Minimum Component
Tachometer Diagram from Figure 21 and Tachometer with Adjustable Zero Speed Voltage
Output suggested by LM2907 datasheet (TEXAS INSTRUMENTS, 2000) on the Typical
Applications part.
54 Chapter 4 Acquisition and Control System
Figure 22 – Tachometer with Adjustable Zero Speed Voltage Output, adapted from (TE-
XAS INSTRUMETS, 2018)
4.4. Rotation Acquisition Channel 55
∙ C1: This capacitor is charged and discharged every cycle by a 180𝜇A typical cur-
rent source. C1 must not be sized lower than 500-pF due to its role in internal
compensation.
∙ R1: Higher of R1 values increase the output voltage for a given frequency, but
too large will degrade the output’s linearity. Because the current pulses are a fixed
magnitude of 180 𝜇A typical, R1 must be high enough to produce the maximum
desired output voltage at maximum input frequency. At maximum input frequency
the pulse train duty cycle is 100%, therefore the average current is 180 𝜇A and R1
= Vo(max) / 180 𝜇A.
∙ C2: This capacitor filters the ripple produced by the current pulses sourced by
the charge pump. Large values reduce the output voltage ripple but increase the
output’s response time to changes in input frequency.
The output voltage (𝑉𝑂 ) can be calculated using Equation 4.1, 𝑉𝐶𝐶 is the supply
voltage and 𝑓𝐼𝑁 the input frequency.
(︃ )︃
𝑉𝐶𝐶 𝐶1 𝑉𝐶𝐶 · 𝑓𝐼𝑁 · 𝐶1
𝑉𝑅𝐼𝑃 𝑃 𝐿𝐸 = · · 1− (4.2)
2 𝐶2 𝐼2
Finally, the last thing to consider is the maximum attainable input frequency,
determined by 𝑉𝐶𝐶 , C1 and 𝐼2 (180uA) in Equation 4.3.
𝐼2
𝑓𝑀 𝐴𝑋 = (4.3)
𝐶1 · 𝑉𝐶𝐶
According to (PNEUS FÁCIL, 2018), the smallest commercial tyre size in terms of
diameter in Brazil is the standard 165/70R13 which has a diameter (D) of 561.2mm and
the one with the biggest diameter (D) is the standard 265/50R20 having 773mm. Using
Equation 4.4 to calculate the overall length of the wheel gives an aproximately length of
1.762m for the smaller tyre and an aproximately length of 2.428m for the bigger one.
𝐶 =𝜋·𝐷 (4.4)
Considering this upper limit speed of 180kph (50 m/s) and the wheel lengths
calculated above (1.762m and 2.2428m) applied it to Equation 4.5:
𝑆𝑝𝑒𝑒𝑑
𝐹 𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 = (4.5)
𝐿𝑒𝑛𝑔𝑡ℎ
We have maximum frequencies of aproximately 28.36Hz (for the smaller tyre stan-
dard (1.762m)) and 20.59Hz (for the bigger tyre standard (2.428m)).
According to (TEXAS INSTRUMENTS, 2000) the input common mode voltage
is equal to 𝑉𝐶𝐶 - 1.5V, and as 𝑉𝐶𝐶 is equal to 5V (check Section 4.11.1.1) the maximum
output voltage 𝑉𝑂 is equal to 3.5V. Considering a safer upper frequency limit of 150Hz
(approximately six times the maximum calculated frequency of 31.52Hz, this multipli-
cation factor will be explained better in Section 6.2.1) applied to Equation 4.1 with the
𝑉𝐶𝐶 = 5𝑉 , the desired offset voltage 𝑉𝑅𝐸𝐹 (1.225V, check Section 4.4.2) and the maximum
output voltage 𝑉𝑂 = 3.5 gives a time constant (R1*C1) of 0.003. Using the recommended
capacitor values from 21 and commercial resistor values it is possible to achieve this time
constant using R1=300kΩ, with this values 𝑉𝑂 can be described by Equation 4.6.
For instance applying the upper frequency value of 28.36Hz calculated previously
will give a 𝑉𝑂 = 1.6504𝑉 .
Figure 23 shows the designed circuit for the speed acquisition channel.
In addition to the components from the circuit Tachometer with Adjustable Zero
Speed Voltage Output on Figure 22, the following features have been added:
∙ TVS Diode: In order to protect the the IC’s input the SMBJ15CA TVS diode
from Littelfuse was added. This diode has a maximum clamping voltage of 24.4V,
the maximum input voltage is ±28V, thus the TVS will help protecting the input
from overvoltages.
4.4. Rotation Acquisition Channel 57
∙ LPF at the output: R74 and C43 form a LPF to that is used to filter any external
post-conversion noise, it has an aproximate frequency of 16Hz.
𝑉𝑆 − 𝑉𝑂𝑈 𝑇
𝑅𝐵𝐼𝐴𝑆 = (4.7)
𝐼𝐿 − 𝐼𝑄
The used 𝑉𝑆 Voltage Supply will be the 5V obtained in the circuit from Figure
50 in Section 4.11.1.1. According to the datasheet, the minimul operating voltage (𝐼𝑄 in
this case) is 100uA. Moreover, 𝑉𝑂𝑈 𝑇 is naturally 1.225V. The minimum bias current (load
current 𝐼𝐿 ) is of 500nA (according to the (TEXAS INSTRUMENTS, 2000)). Hence, any
value of resistance that guarantee that the current that flows through the regulator is less
than 10mA is acceptable.
58 Chapter 4 Acquisition and Control System
Using a 1kΩ resistor a current of 3.775mA can be achieved. Figure 24 shows the
circuit for the 1V2 reference, the capacitors are just a bypass capacitors recommended by
the datasheet.
√ √
𝑁 𝑜𝑖𝑠𝑒(𝑛𝑣 𝐻𝑧) = 4 · 𝐾 · 𝑅 · 𝑇 · 109 (4.8)
60 Chapter 4 Acquisition and Control System
Because the protection circuit includes two equal resistors, whose noise is uncorre-
lated, that is, the two noise sources are independent of each other—the above result must
be multiplied by the square root of 2 (the root sum square of the two noise voltages) and
it is considered as a general rule design to tolerate additional Johnson Noise from 10 to
30% to the amplifier IC (BRYANT et al., 2000). (DUFF; TOWEY, 2010) suggests using
current-limiting resistors of 100Ω, according to the AD8495 datasheet (ANALOG DE-
√
VICES, 2010), the chosen amplifier (AD8495) has a voltage noise density of 32nV 𝐻𝑧.
Combining this resistors noise with the amplifier noise will produce an overall noise of
√
32.485nV 𝐻𝑧, which is just 1.5% above the amplifier’s own noise. Additional protection
can be achieved using TVS to protect the inputs from differential input overvoltage, con-
sidering a bidirectional TVS with a 10V breakdown voltage, the device will theoretically
limit the differential voltage between 10V and -10V, the AD8495 has overvoltage protec-
tion from -25V to 20V when powered with 5V, so this will help protecting the amplifier
inputs.
With the overloads protection done, another important feature to do is to filter
undesired signals in the inputs to avoid them to be amplified later, this is done by filtering
Radio Frequency Interference (RFI), signal lines (specially for low level signals) are quite
susceptible to RF interference (COUNTS; KITCHEN, 2006). Interference that occurs
on both lines are usually reduced by the amplifiers own CMRR (check Section 2.3.5.1),
but only on a limited bandwidth, also the in-amp rectifier cannot filter differential RF
interference. The chosen amplifier (AD8495) has a -3dB bandwidth at 25kHz, (DUFF;
TOWEY, 2010) suggests setting a common-mode cutoff filter frequency at 160kHz (in
order to guarantee the filter will not attenuate signals inside the 25kHz bandwidth). The
standard circuit for the RFI filter is displayed on Figure 27, resistors R and capacitors C𝐶
are used to filter common-mode interference. Capacitor C𝐷 is connected across the bridge
output to reduce any common-mode rejection errors due to the components mismatch,
that way filtering any differential interference. C𝐷 is usually chosen to be ten times larger
than C𝐶 (ANALOG DEVICES, 2010).
The -3dB common-mode bandwidth of this filter from Figure 27 is given by Equa-
tion 4.9 (COUNTS; KITCHEN, 2006).
1
𝐵𝑊𝐶𝑀 = (4.9)
2 · 𝜋 · 𝑅 · 𝐶𝐶
The -3dB differential bandwidth of this filter from Figure 27 is given by Equation
4.10 (COUNTS; KITCHEN, 2006).
1
𝐵𝑊𝐷𝐼𝐹 𝐹 = (4.10)
2 · 𝜋 · 𝑅 · (2 · 𝐶𝐶 + 𝐶𝑑 )
Using the values for the current limiting resistors (100Ω), Equation 4.9 and the
4.5. Temperature Acquisition Channel 61
Figure 27 – RFI
Besides the RFI filter and the amplifier, only a few components were added. First
a TVS diode with a standoff voltage of 5V was placed on each of the thermocouple
signal tracks. According to (ANALOG DEVICES, 2010), in order to implement sensor
disconnection detection a 1M resistor should be added to the the thermocouple negative
track. More over a post amplification LPF with a cutoff frequency of aproximately 16Hz
was placed in order to filter any remain of noise uncorrelated to the thermocouple signal.
62 Chapter 4 Acquisition and Control System
60𝑘Ω
𝑅𝐺 = (4.11)
𝐺−4
Taking into account the sensitivity of 2mV/V, this means that if the cell is excited
with 2V5, its output will vary from 0 to 5mV.
Since the analog input of the chosen microcontroller (Atmega32U4) is 0 to 5V, we
may to amplify the cell output signal by a factor of 500 in order to use the most number
of bits from de MCU’s ADC. Using Equation 4.11, to obtain a gain of amplification ratio
of the ideal 𝑅𝐺 would be 121Ω. This resistor will cause the IC output to range from
about 0V to 5V (0 to 2.5V for contraction and 2.5V to 5V for tension), using 100% of the
resolution of the microcontroller input.
64 Chapter 4 Acquisition and Control System
Figure 31 shows the schematic of the load cell conditioning circuit with the INA125.
The main components of this circut are
∙ TVS Diodes: INA125 only has overvoltage protection and not ESD protection, so
the TVS Diodes were add to enhance the ESD protection.
∙ Capacitors C18 and C19: Bypass capacitors for the signal lines.
∙ R39 and R43: Jumpers that can be replaced by resistors in case a RFI filter is meant
to be implemented using C24, C26 and C27.
∙ R44, R40 and C16: R44 is used to set the impedance from the load cell signal. R40
and C16 are used to form a LPF with a cutoff frequency of 15.91Hz to filter any
remaining noise from the load cell signal.
4.7. Vibration Acquisition Channel 65
∙ ±3g measurements.
When powered up with 3V3 the sensor has a linear voltage output of 0V for -3g
and 3V3 for 3g. This sensor is ideal for this project, the only issue is the installation.
66 Chapter 4 Acquisition and Control System
The sensor comes in a 16-LFCSP IC (Figure 32), and this IC needs to be installed where
acceleration is intended to be measured, so an additional board is necessary to install it.
The are already embedded solutions that can solve this issue, such as the Adafruit
ADXL335 - 5V ready (ADAFRUIT, 2016). This small (19mm x 19mm) board (Figure
33) has the ADXL335 with the capacitors recommended by the datasheet, with a 3V3
voltage regulator so the board can be powererd up with 5V and with the mounting holes
to fix the board in any surface, making it ideal for this project.
The circuit is composed only of a protection and filtering circuit. TZ16 is a TVS
with a reverse standoff voltage of 5V, R105 and C63 form a LPF with cutoff frequency of
aproximately 1.6kHz, which should not interfere with the passband signal.
∙ First way is the deceleration of the rotor with the aid of the brake system using the
brake pads.
∙ Second way of controlling it by accelerating the rotor with the aid of the electric
motor.
The electric motor used is an alternate current three-phase type, it in order to control it’s
speed a frequency inverter was used. The frequency inverter has many different controlling
interfaces that lead to manipulating the three-phase electric system in order to control
the engine’s rotationary speed.
In this project the frequency inverter used was the Weg CFW-08 (shown in Figure
36, according to (WEG, 2009) there are many port interfaces that can be used to control
the desired speed remotely. One of this ways is using one of the device’s 0-10V analog
input with 8-bit resolution. According to the datasheet, this analog input can be used as
a speed reference with programmable gain (0 to 20), the input parameters are the lower
limit speed (when the input is on 0V) and the upper limit speed (when the input is at
10V).
The issue is that this project is microcontrolled, and MCU’s do not have analog
outputs, so a DAC circuit will need to be implemented. The MCU output will be a PWM
signal, moreover as mentioned in Section 2.3.6.3, a PWM signal can encode information
on a pulse train signal and this information can be decoded using a LPF, i.e. performing
a digital to analog conversion.
𝑉𝑚𝑎𝑥
Δ𝑉 = (4.12)
2𝑛 − 1
As mentioned in Section 4.10, the selected MCU has a five volts high logic level,
this voltage level will need amplification in order to best suit the zero to ten volts input
from the frequency inverter. However, as alrealdy mentioned in Section 4.8.1 the analog
input of the frequency inverter already has a 0-20 programmable gain that will be set to
two.
The maximum ripple is the Δ𝑉 from Equation 4.12. Equation 4.13 (METIVIER,
𝑑𝐵
2013) gives the necessary 𝑑𝑒𝑐𝑎𝑑𝑒 attenuation to guarantee a PWM converted signal desired
ripple.
𝑉𝑅𝐼𝑃 𝑃 𝐿𝐸
(︂ )︂
𝐴𝑑𝐵 = 20 · 𝑙𝑜𝑔 (4.13)
𝑉𝑃 𝑊 𝑀
And Equation 4.15 (METIVIER, 2013), is used to calculate the maximum needed
cuttoff frequency for the further to be designed LPF (Low-Pass-Filter) in order to convert
the PWM signal to an analog voltage. The slope value is the filter slope and for first order
70 Chapter 4 Acquisition and Control System
filters and second order filters this slope value is equal respectively to 20dB/decade and
40dB/decade (METIVIER, 2013).
𝐴𝑑𝐵
𝑓𝑐 = 𝑓𝑃 𝑊 𝑀 · 10− 𝑆𝑙𝑜𝑝𝑒 (4.15)
It is possible to combine all this equations into one single equation to calculate
the needed cutoff frequency, is this done in Equation 4.16.
𝐴𝑑𝐵
𝑓𝑐 = 𝑓𝑃 𝑊 𝑀 · 10− 𝑆𝑙𝑜𝑝𝑒
−20·𝑙𝑜𝑔 (2𝑛 −1)
𝑓𝑐 = 𝑓𝑃 𝑊 𝑀 · 10− 𝑆𝑙𝑜𝑝𝑒
20
𝑛 −1) 𝑆𝑙𝑜𝑝𝑒
𝑓𝑐 = 𝑓𝑃 𝑊 𝑀 · 10𝑙𝑜𝑔(2
𝐾𝑛𝑜𝑤𝑖𝑛𝑔 : (4.16)
(︁ )︁
𝑋
𝑥 · 𝑙𝑜𝑔 (𝐴) = 𝑙𝑜𝑔 𝐴
10𝑙𝑜𝑔(𝐴) = 𝐴
𝑆𝑜 :
20·𝑛
𝑓𝑐 = 𝑓𝑃 𝑊 𝑀 · 2 𝑆𝑙𝑜𝑝𝑒
The default PWM frequency for the defined pins are 𝑓𝑃 𝑊 𝑀 = 490𝐻𝑧, also 𝑛 = 8
(number of bits of the ADC from the frequency inverter) and as according to (METIVIER,
2013), a second order LPF is better for converting a PWM signal to voltage and it has by
default 𝑆𝑙𝑜𝑝𝑒 = −40𝑑𝐵/𝑑𝑒𝑐𝑎𝑑𝑒, using Equation 4.16 it is possible to calculate a maximum
cutoff frequency of 30.68Hz.
1
𝑓𝑐 = √ (4.17)
2 · 𝜋 𝑅1 · 𝑅2 · 𝐶1 · 𝐶2
Figure 39 shows the input and output voltage for 25% configured duty cycle, after
the settling time the average output voltage was 1.19V.
Figure 40 shows the input and output voltage for 50% configured duty cycle, after
the settling time the average output voltage was 2.41V.
Figure 41 shows the input and output voltage for 75% configured duty cycle, after
the settling time the average output voltage was 3.60V.
Overall the simulation confirmed the calculations, the average error of the output
in respect with the theoretical values of of 4%.
Item 9 from Section 4.2 states that the system must have a digital output channel
to control a relay). According to (SONGLE, 2018), a standard 5V relay will have a nominal
current of 89.3mA, the microcontroller’s datasheet (MICROCHIP, 2018) says that the
maximum DC current per I/O pin is 40mA though. Hence it is not possible to activate a
relay connected directly to a MCU’s I/O pin.
In order to solve this a relay driver circuit needs to be used, in this case a low side
mosfet driver similar to the one from Figure (ELECTRONICS TUTORIALS, 2018) will
be used.
This is a low-side driver, when the PMOS is submitted to a high logic level (5V)
in it’s gate it will enter on the cutoff region and will not conduct current, hence not
activating the relay. On the other hand, when a low logic level (0V) is applied to the
74 Chapter 4 Acquisition and Control System
device’s gate it will enter the saturation region and will conduct current, hence activating
the relay.
The digital output circuit will be very similar to the one from Figure 43. The
power supply connected to the mosfet will be the 5V source (check Section 4.11.1.1), on
the PMOS drain there will be a TVS diode (SMBJ12A from Littlefuse (LITTLEFUSE,
2015)) (check Section 2.3.7.2), this TVS will help protecting the PMOS from reverse
current and from overvoltages. Figure 44 show the equivalent circuit for the digital output.
The selected PMOS is the NDS332P from Fairchild (FAIRCHILD, 1997), it has
a maximum operating continuos current of 1A, maximum gate threshold voltage of 1V,
maximum drain-source voltage of 20V and maximum gate source voltage of ±8V. It is
adequate for this purpose. Resistor R7 is used to limit the magnitude of the current spike
that can occur when a driver switches fast and has to charge or discharge a large gate
capacitance. Resistor R5 is a pull-up for the gate, it will ensure that the PMOS will be
in cutoff region if there is no input for the MCU. Resistor R8 is optional, for activating
relays it is not needed, but other loads may require a pull-down, that is why it was placed
on the PMOS drain. It was decided to keep any relay out of the board to save space.
Moreover, keeping the relay out of the board gives room for using this digital output for
other purposes. Although the project Requirements only specifies two digital ouputs, as
4.10. MCU 75
there were unused MCU digital ports, a third digital output will also be placed on the
PCB layout.
4.10 MCU
The chosen MCU for the system is the ATMEGA32U4 manufactured by Microchip.
According to (MICROCHIP, 2017a) and (MICROCHIP, 2018) the main features of this
76 Chapter 4 Acquisition and Control System
MCU are:
∙ C28 and C29 : This are bypass capacitors recommended by the datasheet.
∙ S1, D7 and R60 : RST is the reset pin of the MCU, it resets the MCU when low
logic voltage is detected. S1 is a push-button used to short circuit the RST pin to
ground and reset the MCU. R60 is a pull-resistor used to guarantee that RST pin
will remain in high logic voltage when the push-button is not being pressed.
78 Chapter 4 Acquisition and Control System
∙ L1 and C32 : This components form a LPF for the AVCC pin, which is the power
supply for the MCUs ADC.
∙ C33 and C34 : Both are bypassing capacitors recommended by the datasheet, C33
for the ADC internal voltage supply and C34 for the USB 2.0 controller.
∙ SCK, MOSI and MISO: Those are signal used to program the MCU through ISP(In
System Programming).
∙ XTAL1 and XTAL2 : Respectively the input and output for the crystal oscilator
circuit.
∙ VUSB: This is the suppy voltage pin for the USB controller, it is being powered by
an external USB host connected to the board.
∙ RXLED and TXLED: Pins that goes to low logic level when data is being trans-
mitted and high logic level when not.
4.11. Voltage Supplies 79
The USB Port Protection circuit was entirely based on the circuit from Arduino
Leonardo, the only component specified by the MCU datasheet are the 22Ω series resistors
as shown in Figure 49.
L2 and L3 are ferrite beads used to filter high frequency noise. Z1 and Z2 are
ESD supressors, alongside with the series resistors used to protect the MCU’s USB data
lines. Finally, C35 and C36 are bypass capacitors used to stabilize the voltage on the USB
power line.
input voltage supply. There are many different types of voltage regulators, nowadays the
most common DC/DC being switching regulators. They are more efficient than linear
regulators and consequently they dissipate less heat, the downside is that they tend to
have some ripple at the output (SCHWEBER, 2017). However, as the ripple can easily
be filtered and that a switching power supply allows the input voltage to be of a much
bigger range, this solution was considered the best for the project.
4.11.1.1 5V Supply
The voltage supply was chosen to be a 5V DC supply because all the components
can work with this voltage. The chosen voltage regulator for the 5V supply was the
TL2575-05 from Texas Instruments (TEXAS INSTURMENTS, 2014), it has an output
up to 1A, voltage drop of 2V, typical efficiency of 88% and can work with a 7 to 42V
input. Figure 50 shows the circuit used for the 5V supply.
and protect the device. The datasheet recommends using a 3A reverse current diode, a
330uH inductor and 330uF capacitor for the 5V regulator.
The other components of the circuit are:
∙ F1 : resetable fuse, they will limit the current that flows into the circuit and protect
the TVS.
∙ C39 : Another bypass capacitor on the output to filter any residual noise.
As said in Section 4.5.3, the sensor detection circuit needs a 4V1 voltage reference
for the comparator. This is achieved using the LM4040-4.1 from Microchip (MICROCHIP,
2017b). This is a fixed 4V1 precise voltage reference and according to the datasheet it has
a tolerance of ±1.15%. Figure 51 shows the circuit for the 4V1 voltage reference.
According to the datasheet the current flowing through the LM4040-4.1 must be
never be greater than 15mA, voltage reference pins do need very little current, so a third
82 Chapter 4 Acquisition and Control System
of this 15mA will be taken as goal. Considering the voltage drop of 0.9V (5V - 4.1V) and
the current of 5mA we need a 180Ω resistor in series with the regulator.
∙ Instead of soldering the connectors directly to the board, empty pads for soldering
wires were placed, this way connectors can be chosen later and fixed on the board
housing instead of on the board surface.
∙ Protection devices such as TVS, Fuses and current limiting resistors were placed
very close from the terminal pads in order to prevent cooper tracks to be damaged.
∙ SMT (Surface Mount Technology) was a priority during design in order to save
space on the board.
∙ Cooper tracks width: The width of every track was chosen according to the electrical
signal it is carrying.
Figures 52 and 53 shows respectively the top and the bottom side of the board
during design, most of empty spaces were replaced by ground planes in order to make
the board corrosion faster and to make a stronger gnd connection. Other interesting fact
is the width of power supply tracks, it is easy to notice they are much more thick than
normal signal tracks. The board dimmensions are 130mm x 80mm.
4.12. Circuit Fabrication and Assembly 83
5 Firmware Project
∙ Setup: This part of the code is only executed once, as the name may indicate, it is
used to set properties, configure timers, inputs and other features of the hardware.
∙ Loop: This part of the code is executed continuously or until some condition is
reached.
∙ Acquisition Timer: Timer max count value is setted and an interruption service
routine is appointed, this timer is used for the analog sampling/reading function.
∙ LED Timer: Same thing done for Acquisition Timeris done for LED Timer, this
timer controls the blinking frequency of the MCU debug LED.
∙ Serial Port: The serial port baud rate is defined and the serial port is opened.
∙ Port mapping: The port mapping is done, i.e. digital I/O ports are defined as inputs
(high impedance) or as outputs (low impedance) and the analog ports are setted.
After setting up, the code enters in a state in which it waits for a command that can
be from three different groups. This commands are coded into ASCII characters (ASCII,
2017). Table 2 shows all proposed commands.
Each group of commands will lead to a specific block described in Figure 56.
86 Chapter 5 Firmware Project
∙ Digital Output Control: As mentioned in Section 4.3.3.1, there are three digital
outputs on the circuit board. Hence, there will be two commands for each digital
channel (on and off) plus a command to turn all the digital outputs off.
∙ Analog Output Control: Commands from this block will set the duty cycle for the
MCU PWM outputs that will be converted to analog outputs, as explained in Section
4.8.
∙ Acquisition State: There is only one command that will lead the code to this loop
(Filter Read). When in this loop, the code will read the analog outputs for a specific
number of times (Number Of Samples), i.e. sampling the signals. Each sample is
summed up and when the desired number of samples is reached, the code divides
the sum of all samples by this number, giving an averaged read. Then, this read is
print together will the digital inputs read to the upstream data port.
Additional to this commands there are the commands Print Version and One
Read, the first is a command that prints the firmware version, the latter performs one
read of the analog inputs and is only used for debugging.
5.2. Code Map 87
6 Experimental Tests
After the board assembly and fabrication phases were finished to ensure that this
project attends its requirements. This chapter presents the experimental procedure, the
results as well with the discussion and possible improvements. This chapter will be splitted
in two parts, Section 6.1, Hardware Interfaces Test, will deal with the test of each acqui-
sition channel and with the two analog outputs. Section 6.2, Brake Test, deals with the
tests with the materials from Laboratory of Wear of Materials, this tests aim to validate
the hardware solution as a whole.
The experimental procedure is based on the brake test procedures proposed by
(SAE, 2003) described in Section 2.2.2. Basically the test will compreheend turning the
electric motor on (with the aid of the frequency inverter) and by monitoring the Speed
Acquisition Channel (check Section 4.4) wait until the rotor frequency reaches a pre-
defined upper limit and then after a configured delay time, stop accelerating the motor and
starts to brake until the system reaches a lower speed limit. During the entire procedure
the brake pressure, and the temperature on the disc shall be measured. This procedure
should be repeated for a set number of times called snubs.
∙ Waveform = Triangular
A triangular waveform was chosen because it is the most similar to a CKP signal as
Figure 4 shows. For the tests the input frequency was varried from nought to 150Hz
(check Section 4.4.1.2) at a 10Hz step.
The theoretical voltage output was calculate at Section 4.4.1.2 and is described by
Equation 6.1 bellow.
𝑉
𝑉𝑂 = 1.225𝑉 + 0.015 · 𝑓𝐼𝑁 · (6.1)
𝐻𝑧
The average voltage gain from the circuit of 13.68Hz and the experimental error
was 8.8%, the standard deviation was only 0.01 Hz though. That means that even though
the circuit gain is almost 10% different from the theoretical gain, the circuit gain varries
very little along the frequency input range.
The major difference the calculated to the real voltage output Equation is the
voltage offset. Although it was designed to be 1.225V, the measured value was 0.831V,
this is probably due to the lack of a high impedance buffer between the voltage reference
6.1. Hardware Interfaces Test 91
circuit and the frequency-to-voltage converter offset pin. A future version of this project
should consider an operational amplifier to fix this issue. However, as Table 3 shows, the
circuit has a linear output in respect to the input, the only difference during the tests is
that test applications should consider a 0.831V offset instead of a 1.225V offset.
The voltage output of the Rotation Acquisition Channel can be calculated using
Equation 6.2 bellow.
𝑉
𝑉𝑂 = 0.831𝑉 + 0.01368 · 𝑓𝐼𝑁 · (6.2)
𝐻𝑧
6.1.2.2 Results
6.1.2.2.1 Channel 1
Table 4 shows the results from the test described above done at the first Tempe-
rature Acquisition Channel.
The average error from the circuit from the thermocouple tip to the value measured
in the microcontroller’s analog-to-digital-converter is 3.02% and the standard deviation
is 1.25%.
6.1.2.2.2 Channel 2
Table 5 shows the results from the test described above done at the second Tem-
perature Acquisition Channel.
92 Chapter 6 Experimental Tests
The average error from the circuit from the thermocouple tip to the value measured
in the microcontroller’s analog-to-digital-converter is 2.27% and the standard deviation
is 1.06%.
6.1.2.3 Discussion
voltages from -4mv to 4mV at a 0.5mV step. This input voltage range was chosen because
according the amplifier datasheet (INSTRUMENTS, 1997), when the device is powered
up with 5V the voltage output can only typically go from 0.3V to 4.2V, and this input
voltage range amplified by a factor of 500 with an offset of 2.5V will not go beyond the
specified output voltage range.
6.1.3.2 Results
6.1.3.2.1 Channel 1
Table 6 shows the results from the test described above done at the first Brake
Force Acquisition Channel.
The average experimental error for this amplification circuit was 2.43% and the
standard deviation is 1.81%.
6.1.3.2.2 Channel 2
Table 7 shows the results from the test described above done at the second Brake
Force Acquisition Channel.
The average experimental error for this amplification circuit was 1.92% and the
standard deviation is 1.39%.
94 Chapter 6 Experimental Tests
6.1.3.3 Discussion
With this test the Brake Force Acquisition Channel amplifier showed satisfatory
results, with errors of less than 3%. The combination of this precise amplifier and the
Precision Voltage Reference provided by this IC gives a great compound solution for
dealing with load cell signals.
The electric motor used is manufactured by WEG, this is a three-phase motor that
has 12.5 cv of power, 1755 max RPM and has an efficiency of 87.7% and it is connected
to the brake disc machine through a pulley system. Figure 59 the motor itself.
The Load Cell used to measure the brake force has a sensibility of 2mV/V and a
maximum load of 500kgf, meaning when it is powered with 2V5 (check Section 4.6.2) and
submitted to its maximum load the voltage output will be of 5mV. Figure 63 shows how
the load cell can be used to measured the braking force and Figure 64 shows the actual
load cell installed on the Brake Disc Machine.
This project uses two generic type K themocouples at each end of the braking disc,
as Figure 65 shows.
6.2. Brake Test 97
This project also includes a generic Crankshaft Position Sensor, to enhance the
CKP signal amplitude six small magnets were placed around the inertial disc that the
CKP is perpendicular mounted to, the only important consideration when doing this is
that the CKP will actualy measure a frequency six times higher than the actual rotation
frequency of the axel. Figure 66 shows how the CKP sensor was placed.
Additional materials used for the test are the developed hardware and the laptop
used to run the test.
The main screen is divided in three parts, first part displays the sensors readings,
thermocouples measured temperature, rotation frequency and RPMs and load cell braking
force. On the bottom of the screen the test configurations, them being:
∙ Number of Snubs: The number of times each brake iteration will be done.
∙ Interval between snubs (s): The amount of seconds that an interface will wait at the
end of a brake iteration before accelerating the motor again.
∙ Upper Limit (RPM): At each snub the system will accelerate the electric engine
until the rotor reaches this pre-defined rotation rhythm.
∙ Upper Wait Interval (s): After reaching the upper rotation limit the system will
wait for this interval before braking.
∙ Lower Limit (RPM): The system will perform braking until the rotor reaches this
lower RPM limit.
∙ Lower Wait Interval (s): After reaching the Lower Limit the system will for this
interval.
The software itself will take full control of the test, it will turn the electric motor
on when needed, turn the brakes one
showing that the measured quantities are directly related to the tests are being carried
out. Hence, it is expected that temperature should increase after each snub. Morover, as
the measured braking pressure increases and the CKP signal frequency should decrease
rapidly. Moreover, it shall be possible to see a variation in brake efficiency when the brakes
starts to get hotter.
Based on experiment results, it is possible to see the correlation between the mea-
sured quantities. Clearly the measured rotation falls rapidly as the brake force saturates to
a maximum brake force value. Furthermore, at each brake event, detected when the brake
force increases, the temperature measured on each of thoose thermocouples increases.
Taking a more detailed look at the temperature variation at each snub, Figure 69
show the measured brake force regard the measured temperature at the first ten snubs.
This graph states the correlation between the temperature of the brake pads and the
braking action, and the fact that the temperature increases during the braking act is
perfectly in agreement with the fact that a braking system can be analysed as a kinectic
to thermal energy converter, at is was explained in Section 2.1.
Whereas the temperature increases at each snub, there may be a point where the
temperature reaches an equilibrium, Figure 70 shows the variation of temperature during
102 Chapter 6 Experimental Tests
Figure 69 – Temperature and Brake Force Measurements of The First Ten Snubs
the 100 snubs, it is possible to see that even that the temperature bounces at each snub,
there may be a point were the temperature stabilizes close to a certain temperature.
Another interesting point to analyse is the effect of the temperature of the pads
and the rate of desaceleration of the rotor. Figure 71 shows the desaceleration curves at
each snub and shows that during the test the brake efficiency was slightly reduced as the
temperature would increase.
6.2. Brake Test 103
In order to test the analog output of the system an additional test using just
a serial terminal was performed to validade the analog output and its usage with the
frequency inverter. Before starting this test the frequency inverter was configured to take
the 0-10V analog input as its speed reference with a gain of two and maximum rotation
of 500rpm. Whereas, when 5V is present at this analog input the rotor should spin at a
rotation close to 500rpm and when 2.5V is present at the input the rotor should spin at
a rotation close to 250rpm.
The test was carried out varying the analog output from 50 to 100% duty cycle
(2.5 to 5V), taking 50 samples at every 10% interval, i.e. taking 50 samples at 50%, 50
samples at 60%, 50 samples at 70%, 50 samples at 80%, 50 samples at 90% and finally
50 samples at 100%. At each sample the CKP acquisition channel and the analog output
feedback channel were monitored.
Figure 72 shows the analog output voltage in respect with the duty cycle.
It is possible to see that the measured output was really faithful to the theoretically
calculated output, the average error was just 3.35% which is pretty low. If we disconsider
the measured values from 0% and 100% this erros falls to 1.58%, probably this is related
to the fact that 0% and 100% produces a theoretically output of 0V and 5V, which is
104 Chapter 6 Experimental Tests
respectively the negative and positive supply of the operational amplifier from the analog
output circuit (check Section 4.8.5). Although the error for 0% and 100% is more than
two times the error from the rest of the possible duty cycles, it still is quite small, and
this happens because the chosen amplifier for the analog output circuit has rail-to-rail
input/output.
Different from the brake tests from Sections 6.2.3 and 6.2.4, the frequency inverter
was always on, the consequence of this is that the noise from the electric motor and the
frequency inverter was always interfering with the CKP sensor signal. Figure 73 shows
the measured rotation in respect with the duty cycle.
The signal is extremely noisy, compared to the theoretical rotation. However, even
with all that noise it is possible to see that the signal is varying within a range close to
the desired rotation. Maybe with a more sophisticated software it would be possible to
treat this signal better to achieve better results.
In order to test this interface, a Full Stop Test similar to the one from Section
6.2.3 was performed. However, with only 10 snubs and this time, only the rotor spinning
frequency and the accelerometer signal were monitored. Moreover, the rotor spinning
frequency was also monitored when the machine was stabilizing at the upper speed limit.
6.2.7 Results
Figure 74 shows the results from the vibration test, each of the orange pikes re-
present a snub. Moreover, it is important to state that mg means 10− 3 · 9.8 𝑠𝑚2 and not
miligrams.
6.2.7.1 Discussion
The goal of this tests was to show how the acceleration perpendicular to the shaft
would vary along a testing snub. From the results it is possible to take the following
conclusions:
∙ The acceleration, i.e. vibration of the machine is greater when the rotor is spinning.
The average acceleration during this phase was 336.73 mg.
106 Chapter 6 Experimental Tests
∙ At each snub, when the brake phase starts, i.e. when the rotation starts to decrease,
vibration is minimal, almost none, probably noise has a great contribuition to this
signal. The average acceleration during the braking phase was 23.67 mg.
∙ During the act of release of the brakes, when the rotor reach the lower speed limit,
there is a small peak of vibration.
∙ When the brake is released and the motor is turned off, the vibration is quite small.
The average acceleration during the this phase was 40.80 mg.
It was also clearly that the vibration range of the chosen sensor,-3g/+3g (DEVI-
CES, 2010) was too wide for the measurements, the absolute maximum measured accele-
ration value was 656 mg (or 0.652g). A sensor with a narrow measurement range would
better suit this application.
107
7.1 Conclusions
The goals of this work can be summarized as being the project of an electronic
test bench to perform brake tests in automotive components, and this goal was achieved.
As the tests proved the testbench can perform the required data acquisition of the
specified physical quantities listed on Section 4.1 and also meet the mandatory require-
ments from Section 4.2 as well as the system can control the actuators needed for the
brake tests.
During the project phase of this system, versatility was always in mind, the system
does not only have the basical functionalities specified on SAEJ2522, it also achieved all
the optional requirements from Section 4.2. At the end, a multifunctional hardware with
transient protection in all ports was delivered. An interesting point of the developed
hardware is that as it operates with pre-defined commands comming from its generic
USB emulated serial port, any higher level software can theoretically be used to control
it, as long as it respects the defined communication protocol.
The results of the brake tests performed shows a crucial point of this system,
repeatability, the braketests were carried out for a hundred snubs without any interruption
or bug, making the system extremely reliable. The transient protections circuits worked
really well, even in the noisy and transient-dangerous environment created by the three-
phase motor and frequency-inverter, not a single circuit was damaged during the tests.
As mentioned before, this system is extremely versatile, making it able to perform
many different functions. Although this system was initially developed with the final goal
of performing braketests, at another point of view it can be seen as a control and acqui-
sition system. For example, considering it has thermocouple reading inputs and digital
inputs, it could be used to control a SMT (Surface Mount Technology) reflow oven tem-
perature. Moreover the force acquistion channels can be used to control an industrial
precision scale. Considering a vibration channel was implemented, even a vibration test-
bench could be controlled and monitored. Possibilities are infinite, the only restriction is
the amount of inputs and outputs of the designed hardware.
could make the system more capable of being readily changed and maybe developing a
system using Java could even make it easier to migrate the system to a web plataform.
A future version of this project could maybe consider switching from an integrated
architecture to a distribuited one, instead of putting all acquisition channels in one board,
a single board for each acquisition channel could be created and all this boards could
communicate through an I2C or RS485 bus for example. That way the system versatility
would be only limit to the number of allowed devices on the bus. Moreover, as the sensors
cables could be smaller (peripheral boards would be placed closer to the sensors), there
would probably be less noise captured by those cables.
109
References
ARDUINO. Arduino Leonardo Schematic. 2010. Cited 2 times in page 77 and 79.
BARTZ, M.; ZHAKSILIKOV, M.; OGAMI, K. Y. Data driven method and system for
monitoring hardware resource usage for programming an electronic device. [S.l.]: Google
Patents, 2004. US Patent 6,785,881. Cited in page 40.
CNW. Volkswagen Gol 1.0 Ficha Técnica. 2017. Disponível em: <http://www.
carrosnaweb.com.br/fichadetalhe.asp?codigo=1528>. Cited in page 25.
CNW. Volkswagen Gol Plus 1.0 Ficha Técnica. 2017. Disponível em: <http:
//www.carrosnaweb.com.br/fichadetalhe.asp?codigo=431>. Cited in page 25.
o
CONTRAN. Resolução N 519. 2015. Cited in page 26.
DILLON, B. N.; GRIFFEN, N. C.; WEIHS, M. E. Load cell. [S.l.]: Google Patents, 1989.
US Patent 4,815,547. Cited in page 36.
GOODYEAR. Brake Calipers: The Top 3 Things You Need to Know. 2018. Disponível
em: <https://www.goodyearautoservice.com/en-US/brakes/calipers>. Cited 2 times in
page 29 and 31.
GUMS, J. Types of Temperature Sensors. 2018. Disponível em: <https://www.digikey.
com/en/blog/types-of-temperature-sensors>. Cited in page 33.
HALDERMAN, J. D.; MITCHELL, C. D. Automotive brake systems. [S.l.]: Prentice
Hall, 2016. Cited in page 25.
HERNANDES, D. Relembre todos os carros que já lideraram as vendas no Brasil
em 60 anos de hitória. 2017. Disponível em: <https://www.flatout.com.br/
relembre-todos-os-carros-que-ja-lideraram-as-vendas-no-brasil-em-60-anos-de-historia/
>. Cited in page 25.
HOLMES, D. G.; LIPO, T. A. Pulse width modulation for power converters: principles
and practice. [S.l.]: John Wiley & Sons, 2003. v. 18. Cited in page 42.
INC.(NORWOOD, M. A. D. Operational Amplifiers Selection Guide 2011-2012:
Detachable Charts Inside for Quick Product Selection & Design Equations. [S.l.]: Analog
Devices, 2011. Cited in page 38.
INSTRUMENTS, N. Definition of Strain. 2016. Disponível em: <http://www.ni.com/
cms/images/devzone/tut/a/83a1fe69763.gif>. Cited 2 times in page 11 and 35.
INSTRUMENTS, N. Wheatstone Bridge Circuit. 2016. Disponível em: <http:
//www.ni.com/cms/images/devzone/tut/a/a28b553b1069.gif>. Cited 2 times in page
11 and 36.
INSTRUMENTS, T. Instrumentation amplifier. 1997. Cited 4 times in page 11, 63, 64,
and 93.
INSTRUMENTS, T. INA 118. 2000. Cited 2 times in page 11 and 39.
INSTRUMENTS, T. INA118 Precision, Low Power Instrumentation Amplifier. [S.l.],
2000. 29 p. Cited in page 38.
JAMES, C. Fundamentals of linear electronics: Integrated and discrete. Fundamentals of
linear electronics: integrated and discrete. Cengage Learning, 2001. Cited in page 42.
JR, K. N. Equivalence principle for massive bodies. ii. theory. Physical Review, APS,
v. 169, n. 5, p. 1017, 1968. Cited in page 37.
KAWAGUCHI, H. Comparação da análise de conforto de frenagem subjetiva x objetiva
de um veículo de passeio. São Paulo, Brazil, p. 101, 2005. Cited in page 26.
KINZIE, P. A.; RUBIN, L. G. Thermocouple temperature measurement. Physics Today,
v. 26, p. 52, 1973. Cited in page 35.
LIMPERT, R. Brake design and safety. [S.l.: s.n.], 1999. Cited in page 29.
LITTLEFUSE. Transient Voltage Suppression Diodes, SMBJ Series. [S.l.], 2015. Cited
in page 74.
LITTLELFUSE. ESD Test Wave. 2015. Cited 2 times in page 11 and 43.
112 References
PALLÃ, R.; WEBSTER, J. G. et al. Sensors and signal conditioning. [S.l.]: John Wiley
& Sons, 2012. Cited in page 35.
WALTERS, K. How To Select Transient Voltage Suppressors. [S.l.], 2016. Cited in page
45.
WEG. CFW-08 Frequency Inverter. [S.l.], 2009. Cited 2 times in page 68 and 94.
WINDOW, A. L.; HOLISTER, G. S. et al. Strain gauge technology. [S.l.]: Applied science
publishers, 1982. Cited in page 36.